The Prognostic Significance of 50% Coronary Stenosis in Medically Treated Patients with Coronary Artery Disease

Transcription

1 The Prognostic Significance of 50% Coronary Stenosis in Medically Treated Patients with Coronary Artery Disease PHILLIP J. HARRIS, MB., D. PHIL., VICTOR S. BEHAR, M.D., MARTIN J. CONLEY, M.D., FRANK E. HARRELL, JR., PH.D., KERRY L. LEE, PH.D., ROBERT H. PETER, M.D., YIHONG KONG, M.D., AND ROBERT A. ROSATI, M.D. SUMMARY In this study we determined the prognostic significance of 50% coronary stenosis in 1183 medically treated patients with coronary artery disease. Clinical outcome was measured by survival and eventfree (freedom from death and infarction) rates. Significant disease was first defined as 75% or greater narrowing. In 225 patients with less than 75% narrowing of all vessels, including 68 patients with 50% stenosis of at least one vessel, the 3-year survival rate was 100%. Patients with one, two or three significantly (75% or greater) stenosed vessels with additional 50% stenosed vessels had the same outcome as patients with one, two or three diseased vessels without additional 50% stenosed vessels. Significant disease was then defined as 50% or greater narrowing. Patients with significant (50% or greater) stenosis of one, two and three vessels and the left main coronary artery were divided into those in whom all diseased vessels were 75% or greater stenosed (group A) and those in whom at least one vessel was only 50% stenosed (group B). In every category, group B patients had a better outcome than group A patients. The largest differences were in three-vessel and left main coronary artery disease. Group B patients also had lower prevalences of previous infarction and abnormal ventricular function than group A patients. In three-vessel disease, the differences in outcome between group A and group B patients remained significant in multivariable analyses with descriptors of left ventricular function. Thus, 50% coronary stenosis is associated with less risk than 75% or greater stenosis even after adjustment for left ventricular function. When 50% stenosis is defined as significant, subsets based on the number of diseased vessels may be heterogenous with respect to baseline characteristics and outcome. THE EXTENT of coronary artery involvement is an important determinant of outcome in coronary artery disease. -7 Survival is closely related to the presence or absende of left main coronary artery (LMCA) disease and the number of diseased vessels.1-7 To determine the extent of coronary involvement, the degree of stenosis considered significant must be defined. In clinical studies of coronary disease, some groups have defined 50% narrowing of the luminal diameter as significant,5 8, 9 but others have required 70-% stenosis Evidence supporting these definitions has come from acute experiments in which coronary blood flow or myocardial perfusion has been examined under resting and high-flow conditions.'6-18 There is little information available about the effect of the From the Division of Cardiology, Department of Medicine, and Department of Community and Family Medicine, Duke University Medical Center, Durham, North Carolina. Supported by contract HRA from the Health Resources Administration, HEW, research grant HL from the NHLBI, research grant HS from the National Center for Health Services, training grant LM and grants LM and LM from the National Library of Medicine, and grants from the Prudential Insurance Company of America and the Kaiser Family Foundation. Presented in part at the 28th Annual Scientific Sessions of the American College of Cardiology, Miami Beach, Florida, February Dr. Harris is a Fellow in Applied Health Sciences of the National Health and Medical Research Council of Australia. Address for correspondance: Robert A. Rosati, M.D., Box 3337, Duke University Medical Center, Durham, North Carolina Received July 26, 1979; revision accepted January 24, 19. Circulation 62, No. 2, degree of stenosis on clinical outcome. In a recent study from this laboratory,20 Conley et al. concluded that the presence of 50% LMCA stenosis did not affect 3-year survival in patients with 75% or greater stenosis of the three other major vessels. Studies from the Cleveland Clinic6 have shown a relationship between the degree of stenosis and survival in single-vessel disease. In this study, we sought to determine the effect of 50% coronary stenosis on clinical outcome across all degrees of vessel involvement and the prognostic implications of defining 50% coronary stenosis as significant. Methods Population From January 1973 to January 1978, 2389 patients underwent cardiac catheterization for suspected ischemic heart disease at Duke University Medical Center. This figure does not include patients who had prior revascularization procedures, congenital heart disease, hypertrophic cardiomyopathy or valvular heart disease other than mitral insufficiency thought to be secondary to ischemic heart disease. Of the 2389 patients, 658 had normal coronary arteries and 548 patients with coronary disease were treated surgically. The 1183 patients with coronary disease who were treated medically form the basis of this report. Patient Selection Decisions about catheterization and therapy were made by the individual physicians in conjunction with

2 50%7o CORONARY STENOSIS AND PROGNOSIS/Harris et al. 241 their patients. Most patients were catheterized because of the severity of their symptoms; some were catheterized to determine coronary anatomy and ventricular function. Of the 1183 patients, 1170 had angina and/or a history of myocardial infarction at the time of catheterization. Incapacitating angina was the main indication for surgery, except in patients with LMCA disease. Although some medically treated patients were unsuitable for surgery, most were suitable for surgery but were not operated on after the information acquired at catheterization, the severity of symptoms and the response to treatment were taken into account. Baseline Data The computerized information system used in this study has been described.7 21 In summary, data on baseline characteristics were collected prospectively on each patient at the time of cardiac catheterization. The data set consisted of two types of variables: clinical variables derived from the history, physical examination, chest x-ray and ECG, and catheterization variables, including standard hemodynamic measurements and descriptors of coronary anatomy and ventricular function. Coronary angiograms were performed in multiple left anterior oblique and right anterior oblique projections. Since November 1969, coronary angiograms have been interpreted in conference by at least two of the same three experienced angiographers. Differences of opinion were resolved by measuring the lesion in question. The interpretation of each angiogram was also checked by the staff cardiologist responsible for the care of the patient. Throughout this study, significant coronary disease was routinely defined as 70-75% or greater diameter stenosis of one or more major vessels. Since January 1973, the severity of each lesion in each coronary artery segment has been recorded on a coronary artery diagram.22 Lesions were graded as causing less than 25%, 25%, 50%, 75%, 95% or 100% narrowing of the luminal diameter.22 The degree of stenosis of a long lesion was estimated at its narrowest point. In vessels that contained more than one lesion, the degree of stenosis was that of the most severe lesion. If there was more than one 50% lesion, the vessel was still considered to be only 50% stenosed. The data were analyzed using two different definitions of significant stenosis (75% stenosis and 50% stenosis). The number of significantly diseased vessels was determined from the left anterior descending (LAD), left circumflex (LCA) and right coronary (RCA) arteries. Only vessels in which there was at least one lesion that fell within the definition of significance being tested were counted as significantly stenosed. For example, a patient with a 50% LAD lesion proximal to the first septal perforator, a 50% LCA lesion and a 75% RCA lesion would be classified as having only one-vessel disease when significant stenosis was defined as 75% or greater narrowing but would be classified as having three-vessel disease when significant stenosis was defined as 50% or greater narrowing. Lesions in the anterolateral, posterior left ventricular and marginal branches of the three major coronary arteries were only defined as obstruction of their associated major vessels if the branches were large enough to accept a bypass graft and the narrowing exceeded the worst lesion in the associated major artery. In a left-dominant circulation in which the posterior surface of the heart was supplied by the LCA, significant obstruction of the proximal LCA was counted as two diseased vessels if the obstruction jeopardized blood supply to both the lateral and posterior surfaces of the heart. In this situation, RCA stenosis did not contribute to the count of diseased vessels. Patients with 50% or greater LMCA stenosis were analyzed separately. Follow-up All patients were followed 6 and 12 months after cardiac catheterization and annually thereafter. Follow-up information was obtained by a staff cardiologist during a clinic visit or by a research associate by telephone. Follow-up was 99% complete. At each interval, functional status was ascertained and new hospitalizations were documented. The patient's physician was contacted for further information when a patient had died or if there was any possibility that a hospitalization had been related to an ischemic event. The physician was asked to confirm his diagnosis for each event, and if relevant, to provide discharge summaries and documentation of each cardiac event. Fatal and nonfatal myocardial infarctions were documented by the standard electrocardiographic and enzyme criteria. A myocardial infarction was considered to be nonfatal if the patient was discharged from the hospital. Data Analysis Cumulative survival rates from the time of cardiac catheterization were calculated by standard life-table methods.23 The cumulative rate at which patients remained free from both death and nonfatal infarction (event-free rate) was calculated in the same manner as the survival rate by treating both death and nonfatal infarction as terminating events The significance of differences in outcome was tested by the Cox regression model.26 Chi-square statistics for the comparison of outcomes within each anatomic category were obtained by comparing the Cox regression parameters for two groups being compared. This approach produced a comparison of overall outcome rather than an interval-by-interval comparison. The time to the first event was used as the dependent variable in the Cox model analysis of eventfree rates. The significance of differences in the prevalence of baseline characteristics was tested by chi-square analysis or Fisher's exact test if there were fewer than five patients in a cell. To determine the effect of 50% stenosis on outcome, significant disease was first defined as 75% or greater stenosis. Patients with no, one, two or three significantly stenosed vessels were divided into those

3 242 CIRCULATION VOL 62, No 2, AUGUST 19 with no 50% lesions in other vessels, and those with one or more other vessels, including the LMCA, with 50% lesions. Survival and event-free rates were compared. Thus, in patients with no significant disease (no 75% or greater stenosed vessels), patients who had one or more 50% stenosed vessels were compared with those who had no 50% stenosed vessels. Patients who had one significantly diseased vessel and one or more 50% stenosed vessels were compared with those who had one-vessel disease and no additional 50% stenosed vessels. Patients who had two significantly diseased vessels and 50% stenosis of the third vessel were compared with those who had two-vessel disease and no 50% stenosis in the third vessel. Patients who had three-vessel disease and 50% LMCA disease were compared with those who had three-vessel disease and no 50% LMCA disease. The same comparisons were also made excluding patients who had 50% LMCA lesions. Patients with 75% or greater LMCA stenosis were not considered in this analysis. To determine the prognostic implications of defining 50% coronary stenosis as significant, the same population was classified using 50% or greater stenosis to define significant disease. Patients with significant (50% or greater) stenosis of one, two and three vessels and the LMCA were divided into those in whom all significantly diseased vessels were 75% or more stenosed (group A) and those in whom one or more significantly stenosed vessels were only 50% stenosed (group B). Thus, group B patients with one-vessel disease had 50% stenosis of one vessel, group B patients with two- or three-vessel disease had various combinations of 50% and 75% or greater stenosed vessels and group B patients with LMCA disease had 50% stenosis of the LMCA. Within the one-, two- and three-vessel and LMCA disease categories, survival and event-free rates of group A and group B patients were compared. The baseline comparability of group A and group B was determined by comparing the prevalences of 14 prognostically important characteristics within the one, two and three diseased vessel categories. To determine whether the differences in outcome between group A and group B patients with three-vessel disease could be entirely accounted for by differences in left ventricular function, a multivariable analysis was performed in which descriptors of coronary anatomy and left ventricular function were analyzed jointly using the Cox regression model.26 Results One hundred seventy-one of 1183 patients died from cardiovascular causes and 83 suffered at least one nonfatal infarct. Eight noncardiovascular deaths were excluded from further analysis by withdrawing (censoring) the patients alive at the time of death. Significant Stenosis Defined as 75% or Greater Narrowing The distribution of patients with no, one-, two- and three-vessel and LMCA disease when significant TABLE 1. The Distribution of 1183 Patients According to the Number of Diseased Vessels WVhen Signi,ficant Stenosis Was Defined as 75% or Greater Narrowing Number of vessels with significant (.> 75%) stenosis Left main coronary artery Number of patients stenosis was defined as 75% or greater narrowing is shown in table 1. Patients with no, one, two and three diseased vessels were divided into those with one or more additional vessels with 50% stenosis and those with no additional 50% stenoses (table 2). Of the 225 patients with no significant stenoses, 68 had one or more 50% stenosed vessels and the remaining 157 had no vessels with 50% (or greater) stenosis. Of the 265 patients with one-vessel disease, 93 had additional 50% stenosed vessels; of the 2 patients with twovessel disease, 83 had additional 50% stenosed vessels, and of the 5 patients with three-vessel disease, 39 had additional 50% LMCA stenosis. The distribution of 50% stenosed vessels within the no, one-, two- and three-vessel disease categories is shown in table 3. There were no patients with three 50% stenosed vessels. The cumulative survival rates for patients with one or more additional 50% stenosed vessels and for patients without additional 50% stenosed vessels are compared within the no, one-, two- and three-vessel disease categories in figure 1. Additional 50% stenosed vessels had no effect on cumulative survival rates (p > 0.2) in any category of significant disease. In the 225 patients who had no 75% or greater stenosed vessels, the 3-year survival rate was 100%, regardless of whether they had 50% stenosed vessels. The only death among these 225 patients occurred at 4 years in a patient with 50% stenosis of the LCA. In one-vessel disease, the 3-year survival rates were 95% in patients with additional 50% stenosed vessels and 94% in patients with no additional 50% stenosed vessels. In two-vessel disease, the 3-year survival rate was 85% in TABLE 2. Patients with No, One, Two and Three Significantly Stenosed (> 75%0) Vessels Divided into Those Without Additional 50%lO Stenosed Vessels and Those with Additional 50% Stenosed Vessels Number of vessels with significant (. 75%) stenosis Number of patients without additional 50% stenosed vessels Number of patients with additional 50% stenosed vessels * *The 39 patients with three-vessel disease and an additional 50% stenosed vessel all had 50%C7 left main coronary artery stenosis.

4 50% CORONARY STENOSIS AND PROGNOSIS/Harris et al. 243 TABLE 3. The Distribution of Patients with 30% 'Stenosed Vessels According to the Number of Significantly (> 75%) Stenosed Vessels and According to the Number of Additional 50% Stenosed Vessels Number of vessels with significant ( 75%) stenosis Number of additional 50% stenosed vessels Left main coronary artery Total patients with additional 50% stenosed vessels, including 11 patients with 50% LMCA stenosis, and 90% in patients with no additional 50% stenosed vessels. In three-vessel disease, the difference, still insignificant, was reversed: the 3-year survival rate was 81% in those with additional 50% LMCA stenosis and 75% in those without 50% LMCA stenosis. Cumulative event-free rates are compared in figure 10Or F 1OOr ~ 0 VESSELS '75% p p 066 _ O VESSEL ' 75% p=0.0 [Additional 50% Stenosed Vessels F e----c Present o-o Absent *A O-A )e 2 VESSELS 1 75% p= 0.7 -v J-A S9 e 2,2 _-e- inn- lvu p'p- 3 VESSELS )75% 4-~~~~~~1- A TB 39 3,3 p \4 FIGURE 1. The cumulative survival rates of patients with no, one, two and three significantly (75% or greater) stenosed vessels. Within each category, the patients were divided into those with one or more additional 50% stenosed vessel and those with no additional 50% stenosed vessels. The 39 patients with three-vessel disease and an additional 50% stenosis all had 50% left main coronary artery stenosis. The number ofpatients followed in each interval is shown at the bottom of each panel. tll LLI z LUJ B h loot 0 VESSELS ' 75% 100 e _ - * o 100 p z 0.5 -s b ~ 1 VESSEL U75% p k Additionol 50% Stenosed Vessels Present o-- Absent VESSELS '75% P'O S 3 VESSELS '75% "x p:0.2 -c : T FIGURE 2. The cumulative event-free rates ofpatients with no, one, two and three significantly (75% or greater) stenosed vessels. Within each category, patients were divided into those with one or more additional 50% stenosed vessels and those with no additional 50% stenosed vessels. The patients with three-vessel disease and additional 50% stenosis all had 50% left main coronary artery stenosis. The number ofpatients followed in each interval is the same as in figure Again, there were no significant differences between those with and those without additional 50% stenosed vessels in the one-, two- or three-vessel disease categories. The maximum difference was in threevessel disease, where the 3-year event-free rate was 71% in patients with additional 50% LMCA stenosis compared with 63% in those without additional 50% LMCA stenosis. In the 225 patients who had less than 75% stenosis of all vessels, there was one nonfatal infarct, which occurred in a patient with less than 50% stenosis of all coronary vessels. Eleven of 57 patients who had 50% LMCA stenosis died. Of these 11, four had two other diseased vessels and seven had three other diseased vessels. When the patients with 50% LMCA stenosis were excluded from the no, one- and two-vessel disease categories, the differences in survival between those who did and did not have additional 50% stenosed vessels remained insignificant (fig. 3). The difference in event-free rates also remained insignificant. Because patients with 50% LMCA stenosis were excluded, patients with threevessel disease are not shown in figure 3. Significant Stenosis Defined as 50% or Greater Narrowing The distribution of patients with significant stenosis of no, one, two and three vessels and the LMCA when significant stenosis was defined as 50% or greater

5 244 CIRCULATION VOL 62, No 2, AUGUST 19 v)1 100r F VESSEL ' 75X p-0.4 Additional 50 Stenosed Vessels * - Present O-O Absent 4 _A A S 00 _ VESSELS '75X p a 8M VESSELS '75X -CL P-0.9,A1S TB FIGURE 3. The cumulative survival rates of patients with no, one and two significantly (75% or greater) stenosed vessels. Within each category, patients were divided into those with one or more additional 50% stenosed vessels and those with no additional 50% stenosed vessels, but patients with 50% left main coronary artery stenosis were excluded from the analysis. narrowing is shown in table 4. Patients with one-, two-, three-vessel and LMCA disease were divided into those in whom all diseased vessels were 75% or greater stenosed (group A) and those in whom one or more diseased vessels had only a 50% stenosis as the most severe lesion (group B). Table 5 shows the distribution of patients in group A and group B. Patients in group B who had one-vessel disease had 50% stenosis of one vessel, but patients with two- and three-vessel disease in group B had various combination of 50% stenosed vessels and 75% or greater stenosed vessels. For example, of the 90 patients with three significantly (50% or greater) stenosed vessels, 72 had one 50% stenosed vessel and two 75% or greater stenosed vessels, 18 had two 50% stenosed vessels and one 75% or greater stenosed vessel and none had three 50% stenosed vessels. Patients with LMCA disease were divided into those with 50% TABLE 4. The Distribution of 1183 Patients According to the Number of Diseased Vessels When Significant Stenosis Was Defined as 50% or Greater Narrowing Number of vessels with significant (. 50%) stenosis Left main coronary artery Number of patients TABLE 5. Patients with Significant Stenosis (> 50%) of One, Two and Three Vessels and the Left Main Coronary Artery Divided into Those with All Diseased Vessels 75% or Greater Stenosed (Group A) and Those with One or More Diseased Vessels only 50% Stenosed (Group B) Number of vessels with significant (> 50%) stenosis Left main coronary artery Number of patients with all 75% or greater stenosed vessels (group A) Number of patients with one or more 50% stenosed vessels (group B) LMCA lesions (group B) and those with 75% or greater lesions (group A). Survival The cumulative survival rates of group A and group B patients with significant (50% or greater) stenosis of one, two and three vessels and the LMCA are shown in figure 4. The survival rate of group A patients with :D 100Ft ~ F 10~ VESSEL > 50s -4--*---.- p=0.2 Group A *---- Group B A B VESSELS>50% 1OOt h F P= 0o 0 _ A a ~ 2 VESSELS>50> _-4- ~~ p4- A :B LEFT MAlN 50% FIGURE 4. In thisfigure significant stenosis was defined as 50% or greater narrowing. Within the one-, two- and threevessel and left main coronary artery disease categories, the cumulative survival rates ofpatients in whom all diseased vessels were 75% or greater stenosed (group A) are compared with those in whom one or more of the diseased vessels was only 50% stenosed (group B). Patients with twoand three-vessel disease who were in group B had a mixture of50% stenosed vessels and 75% or greater stenosed vessels.

6 50% CORONARY STENOSIS AND PROGNOSIS/Harris et al. 245 LLJ z uil 100-3a 1 VESSEL >50% p =O.07 0-O-- Group A *- G Group B 3 VESSELS> 50% 1OO0 ~ F k VESSELS > 50 % p.0.3 LEFT MAIN > 50% p-p.oos 1 FIGURE 5. Cumulative event-free rates of patients with significant (50% or greater) stenosis of one, two and three vessels and the left main coronary artery. The patients were divided into those in whom all diseased vessels were 75% or greater stenosed (group A) and those in whom one or more diseased vessels was only 50% stenosed (group B). The number ofpatients followed in each interval is the same as in figure 4. 75% or greater LMCA stenosis was considerably lower (p < 0.01) than that of group B patients with 50% LMCA stenosis (3-year survival rates of 35% and 81%, respectively). In one-, two- and three-vessel disease, the survival rates of group A patients were lower at every interval than those of group B patients. The differences were largest in three-vessel disease (p < 0.01). In one- and two-vessel disease, the differences in survival between group A and group B were not statistically significant. Ischemic Events The event-free rates of group A patients were also consistently lower than those of group B patients (fig. 5). The differences in three-vessel and LMCA disease were highly significant. With nonfatal infarcts included in the outcome, the difference between group A and group B patients with one-vessel disease was marginally significant (p = 0.07). The difference between group A and group B patients with two-vessel disease was not statistically significant. Baseline Characteristics The baseline characteristics of group A and group B patients with one, two and three significantly (50%o or greater) stenosed vessels are compared in table 6. In each category, group B patients had a lower prevalence of previous myocardial infarction, resting electrocardiographic abnormalities and abnormal ventricular function than group A patients. Many of the differences in prevalence were significant. The lower prevalence of abnormal ejection fraction in group B patients was significant (p < 0.05) in every anatomic category. Multivariable Analysis The results of the multivariable Cox model analysis of outcome in patients with three significantly (50% or greater) stenosed vessels are shown in table 7. The presence of one or more 50% stenosed vessels remained a significant (p < 0.05) predictor of a better outcome (survival and events) after adjustment for differences in left ventricular ejection fraction. Because left ventricular ejection fractions were not measured in 67 of the patients with three-vessel disease, the multivariable analysis was repeated using descriptors of left ventricular function available in every patient: abnormal left ventricular contraction, diffusely abnormal left ventricular contraction, left ventricular end-diastolic pressure and arteriovenous oxygen difference. The presence of one or more 50% stenosed vessels remained a significant predictor of an improved outcome after adjustment for all of these descriptors of left ventricular function. Discussion The evidence that 50% coronary artery luminal diameter narrowing is hemodynamically significant comes largely from acute animal experiments in which coronary flow or myocardial perfusion was examined under high-flow conditions.16l19, Under resting conditions, coronary flow is generally not affected until the degree of stenosis approaches 75%*16-19, Parallel observations in man have yielded similar findings. In acute human experiments, myocardial perfusion defects have been shown by radionuclide imaging under high-flow conditions in patients with % or less coronary stenosis Studies during open heart surgery, however, failed to show that occlusion of a bypass graft decreased myocardial perfusion32 or produced reactive hyperemia33 unless the lesion in the bypassed vessel exceeded 75% stenosis. In this study of clinical outcome in relation to the degree of coronary stenosis, 50% coronary artery stenosis alone was less likely to be associated with death and nonfatal infarction than 75% or greater stenosis. In fact, over the 4-year period of this study, 50% coronary stenosis was not shown to cause any increase in risk. For example, patients with 75% or greater stenosis of two vessels and 50% stenosis of the third vessel had survival and event-free rates comparable not to patients with three-vessel disease but to patients with two-vessel disease. Consequently, defining 50% coronary stenosis as significant distinguished groups that were heterogeneous with respect to clinically important baseline characteristics and outcome.

7 246 CIRCULATION VOL 62, No 2, AUGUST 19 TABLE 6. Prevalence (%) of Baseline Characteristics in Group A and Group B Patients with One, Two and Three > 50%c Stenosed Vessels One vessel Two vessel Three vessel Group A Group B Group A Group B Group A Group B (172 pts) (54 pts) (157 pts) (82 pts) (366 pts) (90 pts) Age > 50 years Duration of symptoms > 2 years t Typical angina 39 25* 58 35t 77 63t Progressive chest pain * NYHA class III or IV heart failure it Previous myocardial infarction 45 lit 55 37t 59 50* Ventricular gallop Cardiomegaly on chest x-ray t Q waves on ECG 26 15* t Resting ST-T-wave changes 41 24t t Ejection fraction < % 13 2t 29 15t 41 10t Arteriovenous oxygen Differences > 5 vol% t LV end-diastolic pressure > 17 mm Hg t Diffusely abnormal contraction t *p <0.1. tp < Abbreviations: NYHA = New York Heart Association; LV = left ventricular. The fact that acute studies suggest the opposite conclusion about the importance of 50% lesions is not surprising. Many variables, not all of which are known, influence the relationship between the degree of stenosis and clinical outcome. Among those that have been considered are the length of the coronary lesion34 and the degree of collateralization.35 The strongest argument that lesions causing less than 75% coronary narrowing are not prognostically significant over 3-4 years comes from the group of 225 patients who had coronary disease but no major vessels with 75% or greater stenosis. There was only one death (at 4 y-ears) and one nonfatal infarct in this group, which included 68 patients with 50% stenosis of at least one artery. This group of 225 patients was virtually free from any bias introduced by the exclusion TABLE 7. Results of MulUivariable Cox Model Analysis in Patients with Three Signi,fcantly Stenosed (. 50%) Vessels Adjusted chi square Adjusted chi square for survival for events One or more 50% stenosed vessels Ejection fraction Chi square > 3.84, p < 0.05; chi square > 6.63, p < of surgical patients, because only four patients were excluded from it by selection for surgery. Fifty percent coronary stenosis was also associated with a more benign history preceding cardiac catheterization. The important differences in baseline characteristics between group A and group B patients with the same number of 50% or greater stenosed vessels could be related to previous myocardial infarction and its consequences. Group B patients, those who had at least one 50% stenosed vessel, had lower prevalences of resting ECG changes and the clinical, chest x-ray, ventriculographic and hemodynamic manifestations of impaired left ventricular function. These differences in baseline ventricular function probably contributed to the differences in outcome. However, the multivariable analysis showed that when differences in baseline ventricular function were considered in patients with three 50% or greater stenosed vessels, the effect of 50% stenosed vessels on outcome remained significant. The similar outcomes of group A and group B patients with two diseased (50% or greater) vessels can be explained by the finding that with the 75% definition, the outcomes of patients with one- and two-vessel disease were not different. The majority of group B patients with two-vessel disease did have one 75% or greater stenosed vessel. The better survival of patients with 50% LMCA stenosis compared with 75% or greater LMCA

8 50% CORONARY STENOSIS AND PROGNOSIS/Harris et al. 247 stenosis has been previously reported in a larger population.20 The present study shows that this difference is maintained when nonfatal myocardial infarction is included in the outcome. Although an association between 50% LMCA stenosis and increased risk in patients who had two vessels with 75% or greater stenosis could be inferred from our results, we believe this was a chance relationship. Four of 11 patients with this combination of lesions died. However, 50% LMCA stenosis did not affect the outcome in threevessel disease in this study or in the larger series,20 and in the present study, there were no deaths among the seven patients who had 50% LMCA stenosis and fewer than two other diseased vessels. Over a longer follow-up period, 50% lesions might progress to 75% or greater stenosis and have a significant effect on outcome. In longitudinal angiographic studies of coronary lesions, disease progression has been unpredictable Thus, rather than counting 50% stenosis as significant when first documented, to allow for possible progression, we recommend classifying patients according to the number of 75% or greater stenosed vessels. Future studies of intermediate and long-term outcome in patients with 50% stenosed vessels may more clearly define the pattern of the progression of 50% stenosis over time. General acceptance of the results of this study will depend on the accuracy and reproducibility of the coronary angiography. In a study published from this center involving patients catheterized in a period just before the commencement of this study,38 angiographic estimates of stenosis and autopsy measurements were compared. There was 79% agreement for lesions causing 75% or greater narrowing. The majority of errors was due to underestimation. The accuracy of estimating lesions causing less than 75% stenosis was not examined. In a later study, two angiographers estimated the presence of 75% or greater stenoses with 88% and 89% accuracy in comparison with autopsy findings and with agreement in 86% of 3 segments (unpublished results). These results are comparable to the accuracy and reproducibility of coronary angiography reported from other centers.39 Patients with coronary disease are usually stratified by important prognostic variables. The results of this study indicate that if 50% stenosis is defined as significant, subsets based on the number of diseased vessels may be heterogenous with respect to baseline characteristics and outcome. Additional stratification by left ventricular function may reduce this heterogeneity. However, in this study, the differences in outcome in patients with three-vessel disease, were not entirely accounted for by left ventricular function. Despite varying definitions of significant stenosis, there has been reasonable agreement among the results of recently reported outcome studies The effect on outcome of combining patients with 50% stenosis with patients with 75% or greater stenosis will obviously depend on the relative proportions of the two lesions. Until more information is available about the prevalence of 50% stenosis in other series, our interpretation of this investigation is that studies using different definitions of significant stenosis must be compared with caution. Further, the difference in definition is likely to most affect the comparison of individual subsets, particularly three-vessel disease and LMCA disease. Acknowledgments The authors thank B. McCants and S. Wells for their assistance in the data analysis and Judie Armstrong and Nancy Terry for their diligence in preparing the manuscript. References 1. Friesinger GC, Page EE, Ross RS: Prognostic significance of coronary arteriography. Trans Assoc Am Physicians 83: 78, Oberman A, Jones WB, Riley C, Reeves TJ, Sheffield L, Turner M: Natural history of coronary artery disease. Bull NY Acad Med 48: 1109, Bruschke AVG, Proudfit WL, Sones FM: Progress study of 590 consecutive nonsurgical cases of coronary disease followed 5-9 years. 1. Arteriographic correlations. Circulation 47: 1147, Reeves TJ, Oberman A, Jones WB, Sheffield LT: Natural history of angina pectoris. Am J Cardiol 33: 423, Burggraff GW, Parker JO: Prognosis in coronary artery disease. Angiographic, hemodynamic, and clinical factors. Circulation 51: 146, Proudfit WL, Bruschke AVG, Sones FM: Natural history of obstructive coronary artery disease: ten-year study of 1 nonsurgical cases. Prog Cardiovasc Dis 21: 53, Harris PJ, Harrell FE, Lee KL, Behar VS, Rosati RA: Survival in medically treated coronary artery disease. Circulation : 1259, Vliestra RE, Assad-Morell JL, Frye RL, Elveback LR, Connolly DC, Ritman EL, Pluth JR, Barnhorst DA, Danielson GK, Wallace RB: Survival predictors in coronary artery disease. Medical and surgical comparisons. Mayo Clinic Proc 52: 85, Murphy ML, Hultgren HN, Detre K, Thomsen J, Takaro T: Treatment of chronic stable angina. A preliminary report of survival data of the randomized Veterans Administration Cooperative Study. N Engl J Med 297: 621, Wertheimer M, Liddle HV: Results of direct coronary artery graft reconstruction. Ann Thorac Surg 20: 538, Guinn GA, Mathur VS: Surgical versus medical treatment for stable angina pectoris: prospective randomized study with 1- to 4-year followup. Ann Thorac Surg 22: 526, Unstable Angina Pectoris: National Cooperative Study Group to Compare Medical and Surgical Therapy. 1. Report of protocol and patient population. Am J Cardiol 37: 896, , Lawrie GM, Morris GC, Howell JF, Tredici TD, Chapman DW: Improved survival after 5 years in 1,144 patients after coronary bypass surgery. Am J Cardiol 42: 709, Hammermeister KE, DeRouen TA, Dodge HT: Evidence from a nonrandomized study that coronary surgery prolongs survival in patients with two-vessel coronary disease. Circulation 59: 430, Webster JS, Moberg C, Rincon G: Natural history of severe proximal coronary artery disease as documented by coronary cineangiography. Am J Cardiol 33: 195, Gould KL, Lipscomb K, Hamilton GW: Physiologic basis for assessing critical coronary stenosis. Instantaneous flow response and regional distribution during coronary hyperemia as measures of coronary flow reserve. Am J Cardiol 33: 87, Gould KL, Lipscomb K: Effects of coronary stenosis on coronary flow reserve and resistance. Am J Cardiol 34: 48, Gould KL, Lipscomb K, Calvert C: Compensatory changes of the distal coronary vascular bed during progressive coronary constriction. Circulation 51: 1085, 1975

9 248 CIRCULATION VOL 62, No 2, AUGUST Folts JD, Gallagher K, Rowe GG: Hemodynamic effects of controlled degrees of coronary artery stenosis in short-term and long-term studies in dogs. J Thorac Cardiovasc Surg 73: 722, Conley MJ, Ely RL, Kisslo J, Lee KL, McNeer JF, Rosati RA: The prognostic spectrum of left main disease. Circulation 57: 947, Rosati RA, McNeer JF, Starmer CF, Mittler BS, Morris JJ, Wallace AG: A new information system for medical practice. Arch Intern Med 135: 1017, Hartman CW, Kong Y, Margolis JR, Warren GS, Peter RH, Behar VS, Oldham HN: Aortocoronary bypass surgery: correlation of angiographic, symptomatic and functional improvement at 1 year. Am J Cardiol 37: 352, Cutler SJ, Ederer F: Maximum utilization of the life table method in analyzing survival. J Chronic Dis 8: 699, Peto R, Pike MC, Armitage P, Breslow NE, Cox DR, Howard SV, Mantel N, McPherson K, Peto J, Smith PG: Design and analysis of randomized clinical trials requiring prolonged observation of each patient. II. Analysis and examples. Br J Cancer 35: 1, Ullyot DJ, Wisneski J, Sullivan RW, Gertz EW, Ryan C: The import of coronary artery bypass surgery on late myocardial infarction. J Thorac Cardiovasc Surg 70: 5, Cox DR: Regression models and life tables. J R Stat Soc (series B) 34: 187, Hillis WS, Faulkner SL, Fisher RD, Dean RM, Friesinger GC, Bender HW: Significance of severity of coronary arterial stenosis: metabolic studies and correlation with hyperemic responses. J Surg Res 18: 125, Gould KL: Noninvasive assessment of coronary artery stenoses by myocardial perfusion imaging during pharmacologic coronary vasodilatation. I. Physiologic basis and experimental validation. Am J Cardiol 41: 267, Gould KL, Schelbert HR, Phelps ME, Hoffman EJ: Noninvasive assessment of coronary stenosis with myocardial perfusion imaging during pharmacologic coronary vasodilatation. V. Detection of 47 percent diameter coronary stenosis with intravenous nitrogen-13 ammonia and emission-computer tomography in intact dogs. Am J Cardiol 43: 200, Selwyn A, Fox K, Forse G, Steiner R: Disturbances of regional myocardial perfusion in patients with coronary artery disease. (abstr) Am J Cardiol 43: 432, Albro PC, Gould KL, Westcott RJ, Hamilton GW, Ritchie JL,Williams DL: Noninvasive assessment of coronary stenoses by myocardial imaging during pharmacologic coronary vasodilatation. III. Clinical trial. Am J Cardiol 42: 751, Smith SC, Gorlin R, Herman MV, Taylor WJ, Collins JJ: Myocardial blood flow in man: effects of coronary collateral circulation and coronary bypass surgery. J Clin Invest 51: 2556, Oldham HN, Rembert JC, Greenfield JC, Wechsler AS, Sabiston DC: Intraoperative relationships between aortocoronary bypass graft blood flow, peripheral coronary artery pressure and reactive hyperemia. In Primary and Secondary Angina Pectoris, edited by Maseri A. New York, Grune and Stratton, 1978, pp Feldman RL, Nichols WW, Pepine GJ, Conti RC: Hemodynamic significance of the length of a coronary arterial narrowing. Am J Cardiol 41: 865, Oldham HN, Jones RH, Harris CC, Howe WR, Goodrich JK, Sabiston DC: Intraoperative relationships between regional myocardial distribution of bypass graft flow and the coronary collateral circulation. J Thorac Cardiovasc Surg 77: 32, Rosch J, Antonovic R, Trenouth RS, Rahimtoola SH, Sim DN, Dotter CT: The natural history of coronary artery stenosis. A longitudinal angiographic assessment. Radiology 119: 513, Marchandise B, Bourassa MG, Chaitman BR, Lesperance J: Angiographic evaluation of the natural history of normal coronary arteries and mild atherosclerosis. Am J Cardiol 41: 216, Schwartz JN, Kong Y, Hackel DB, Bartel AG: Comparison of angiographic and postmortem findings in patients with coronary artery disease. Am J Cardiol 36: 174, DeRouen TA, Murray JA, Owen W: Variability in the analysis of coronary arteriograms. Circulation 55: 324, Hutchins GM, Bulkley BH, Ridolfi RL, Griffith LSC, Lohr FT, Piasio MA: Correlation of coronary arteriograms and left ventriculograms with postmortem studies. Circulation 57: 32, 1977